Fracture resistance depends on the shape of the whole bone with respect to loading, the microstructure of both cortical and trabecular bone, and the ultrastructure of the extracellular matrix or the inherent quality of the bone matrix. While advances in clinical imaging provide measurements of cortical thickness, cortical cross-sectional geometry, cortical porosity, trabecular micro-architecture, as well as volumetric mineral density and matrix- bound water (at distal sites), clinical tools are currently not available to assess the contribution of matrix composition and organization to fracture resistance. Raman spectroscopy (RS) is one technology well suited to fill this gap in clinical diagnostics of bone because it is i) relatively inexpensive and safe with rapid acquisition times, ii) sensitive to both the mineral phase and the organic matrix, and iii) easy-to-use in which a hand-held probe acquires the spectra. There are however several obstacles to overcome in order for RS to provide clinically meaningful assessments of bone matrix quality: i) identify the best strategy to acquire the Raman spectra from bone (transcutaneous vs. percutaneous with or without polarization preserving optics) and ii) determine how to analyze the spectra such that Raman measurements can assess fracture resistance (peak ratios vs. sub-band peak ratios vs. full spectrum analysis). Addressing these challenges, the project will assess the ability of RS to predict mechanical properties of human cortical bone and do so with respect to volumetric bone mineral density (vBMD) and age (Aim 1) and will identify the matrix factors that influence Raman spectroscopic properties of bone quality (Aim 2). For Aim 1, Raman spectra will be acquired from the medial side of the tibia mid-shaft near the anterior ridge (shin) using first a spatially offset RS probe, then using a small fiber optic Raman probe after soft tissue removal, and lastly using a confocal Raman instrument that preserves polarization of the light. Next, mechanical specimens will be extracted from each cadaveric mid-shaft as well as the femoral neck and head, imaged by micro-computed tomography to determine vBMD, and then subjected to tensile testing, fracture toughness testing, or compressive fatigue testing. General linear models will be used to determine whether Raman properties help age and vBMD explain the variance in strength, toughness, fracture toughness, and fatigue resistance (i.e., they add value). Based on preliminary work, we expect a sub-peak ratio within the Amide I to be a predictor of fracture resistance. For Aim 2, pieces of bone from the mechanical specimens (away from test region) will be demineralized and subjected to biochemical assays to determine collagen crosslink concentrations, fluorescent advanced glycation end-products (AGEs), percentage of denatured collagen, and degree of collagen solubility. In addition, Raman spectra will be acquired before and after AGE accumulation and fatigue loading using separate bone samples (mineralized). Since the shape of the Amide I band is reflective of the secondary structure of type 1 collagen, we expect sub- peak ratios to be sensitive to degree of AGE crosslinks and the degree to which the collagen is helical.